CN102487523B - User compliant analysis method and device - Google Patents

Abstract

The invention discloses a user compliant analysis method and device, wherein the user compliant analysis method comprises the following steps: collecting the number of complaints of each type of users; making statistics of average compliant probability of the complaints of each type of the users; and calculating the predicted number of the complaints of each type of the users according to the number of the complaints of each type of the users and the average compliant probability of the complaints of each type of the users. According to the user compliant analysis method and device, disclosed by the invention, the collected number of the user complaints is amplified by introducing an amplification coefficient (such as average compliant rate of the complaints of the users, correlation coefficient among the complaints of the different types of the users, the number of the solved complaints of one type of the users and the like), so that the comprehensive situation of network quality can be indicated through the complaints of part of the users. According to the user compliant analysis method and device disclosed by the invention, problems in a network can be made clear through the situation that the complaints of the different types of the users are formed, then the network can be optimized according to the demands of user perception and the pertinency of network optimization can be further improved.

Description

Customer complaint analytical method and device

Technical field

The present invention relates to a kind of network management technology, relate in particular to a kind of customer complaint analytical method and device.

Background technology

Current movement/wireless communication system is carrying increasing business, along with the continuous aggravation of Mobile Communications Market competition, network quality becomes the key factor that strengthens enterprise competitiveness more, Ye Shi operator carries out link the most key in market and business development, and therefore the network optimization has become the emphasis that network operation is safeguarded.

Various Key Performance Indicators in the main network-oriented of the network optimization (KPI, Key PerformanceIndicator), for main businesses such as speeches, subregion, point specialty ensure the normal operation of network; Along with the continuous prosperity of business, the network optimization need to, towards the user experience quality of complicated business, be carried out differentiation, optimize end to end for the service quality of mixed service, ensures that user uses normal operation and the Quality of experience of miscellaneous service.

For operator, can the most directly embody user to network quality perception be exactly customer complaint, customer complaint quantity, classification that operator receives have direct directive significance to the network optimization.The technical scheme of carrying out the network optimization according to customer complaint at present, normally collects, classifies received customer complaint, and carries out the judgement of fault and the feedback of complaint according to the rule of specifying.But the defect of these schemes is:

1, the problem of the just certain customers of customer complaint reflection but not comprehensive problem, for some network coverages less than place may there is situation about cannot complain, and some inconveniences or the not good content of perception in user's use business may do not complained, at present the processing of complaining and corresponding Optimization Work are not considered;

2, common only for up-to-date customer complaint of receiving for the processing such as collection, classification of customer complaint, but do not consider customer complaint in the past, particularly same client is for the complaint repeatedly of same problem, and in fact in a period of time the statistical analysis of customer complaint there are more directive significances for the network optimization.

Summary of the invention

The object of the invention is to, a kind of customer complaint analytical method and device are provided, obtain customer complaint information more comprehensively by the complaint of certain customers.

For achieving the above object, according to an aspect of the present invention, provide a kind of customer complaint analytical method, comprising: the quantity that gathers every class customer complaint;

Add up the average complaint probability of every class customer complaint;

The quantity of estimating that goes out every class customer complaint according to the average complaint probability calculation of the quantity of described every class customer complaint and every class customer complaint is:

T _i=N _i/ C _i, wherein, T _ibe the quantity of estimating of i class customer complaint, N _ifor the quantity of every class customer complaint of collecting, C _ifor the average complaint probability of every class customer complaint.

For achieving the above object, according to another aspect of the present invention, provide a kind of customer complaint analytical method, comprising: the quantity that gathers every class customer complaint;

Calculate the coefficient correlation between dissimilar customer complaint;

What calculate every class customer complaint according to the quantity of described coefficient correlation and every class customer complaint estimates quantity T _ifor:

wherein, R _jiit is the coefficient correlation between the customer complaint of j class and the customer complaint of i class.

Preferably, the method also comprises: further estimate quantity T according to the every class customer complaint of average complaint probability calculation of every class customer complaint _ifor:

wherein, C _ifor the average complaint probability of every class customer complaint.

For achieving the above object, according to another aspect of the present invention, provide a kind of customer complaint analytical method, comprising: be captured in the quantity of every class customer complaint in certain hour section and the wherein quantity of a class customer complaint being resolved within this time period;

What calculate the every class customer complaint of next time period according to the described quantity of a wherein class customer complaint being resolved and the quantity of every class customer complaint estimates quantity T _ifor:

wherein, T _i(t) the customer complaint quantity for collecting in the t time period; S _i(t) quantity being resolved for i class customer complaint in the t time period.

Preferably, the method also comprises: according to the quantity of estimating of the every class customer complaint of the Calculation of correlation factor between dissimilar customer complaint be further:

wherein, R _jiit is the coefficient correlation between the customer complaint of j class and the customer complaint of i class.

More preferably, the method also comprises: according to the quantity of estimating of the every class customer complaint of average complaint probability calculation of every class customer complaint be further:

wherein, C _ifor the average complaint probability of every class customer complaint.

More preferably, the method also comprises: according to the quantity of estimating of the every class customer complaint of average complaint probability calculation of every class customer complaint be further:

wherein, C _ifor the average complaint probability of every class customer complaint.

For achieving the above object, according to another aspect of the present invention, provide a kind of customer complaint analytical equipment, comprising: acquisition module, for gathering the quantity of every class customer complaint;

Statistical module, for adding up the average complaint probability of every class customer complaint;

Computing module, for the quantity of estimating that goes out every class customer complaint according to the average complaint probability calculation of the quantity of described every class customer complaint and every class customer complaint is:

T _i=N _i/ C _i, wherein, T _ibe the quantity of estimating of i class customer complaint, N _ifor the quantity of every class customer complaint of collecting, C _ifor the average complaint probability of every class customer complaint.

For achieving the above object, according to another aspect of the present invention, provide a kind of customer complaint analytical equipment, comprising: acquisition module, for gathering the quantity of every class customer complaint;

The first computing module, for calculating the coefficient correlation between dissimilar customer complaint;

The second computing module, estimates quantity T for what calculate every class customer complaint according to the quantity of described coefficient correlation and every class customer complaint _ifor:

wherein, R _jiit is the coefficient correlation between the customer complaint of j class and the customer complaint of i class.

For achieving the above object, according to another aspect of the present invention, a kind of customer complaint analytical equipment is provided, comprises: the first acquisition module, for the wherein quantity of a class customer complaint that is captured in the quantity of every class customer complaint in certain hour section and is resolved within this time period;

The 3rd computing module, what calculate the every class customer complaint of next time period for the quantity of a wherein class customer complaint that is resolved described in basis and the quantity of every class customer complaint estimates quantity T _ifor:

wherein, S _i(t) quantity being resolved for i class customer complaint in the t time period.

Customer complaint analytical method of the present invention and device, by introducing amplification coefficient (as the quantity of the coefficient correlation between average the rate of complaints of customer complaint, variety classes customer complaint and the wherein class customer complaint that is resolved etc.), the customer complaint quantity having collected is amplified, realize the comprehensive situation that presents network quality by the complaint of certain customers.The present invention is by forming the situation of dissimilar customer complaint, the problem existing in can definite network, and then can be optimized network according to the demand of user awareness, improve the specific aim of the network optimization.

Brief description of the drawings

Fig. 1 is the flow chart of network optimization process of the present invention;

Fig. 2 is the structure chart of customer complaint analytical equipment embodiment of the present invention;

Fig. 3 is the structure chart of another embodiment of customer complaint analytical equipment of the present invention;

Fig. 4 is the customer complaint analytical equipment of the present invention structure chart of an embodiment again.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in detail.

The present invention mainly, on customer complaint basis of classification, reasonably amplifies customer complaint.Suppose in section, to have collected the customer complaint of M class sometime in certain region at present, it is (N that every class is complained the actual quantity of collecting ₁, N ₂..., N _m).Customer complaint analytical method of the present invention mainly contains following several mode:

One, be (C according to average the rate of complaints (being the probability that user complains in the time that such problem occurs) of user being investigated to every class customer complaint of statistics ₁, C ₂..., C _m), 0 < C _i≤ 1,0 < i≤M and be integer;

According to the quantity N of every class customer complaint _iand the average complaint probability C of every class customer complaint _ithe quantity of estimating that calculates the customer complaint of i class is:

T _i＝N _i/C _i。

Two, owing to having correlation between certain customers' complaint, for example in fact may imply to the complaint of call drop in communication process the bad problem that covers, calculate the coefficient correlation between dissimilar complaint by Delphi method (being expert's scoring), as the coefficient R between the customer complaint of j class and the customer complaint of i class _ji, wherein, 0 < R _ji≤ 1,0 < i, j≤M and be integer, R in the time of j=i _ji=1, consider the reciprocity that two classes complain coefficient correlations to exist, the use that should reduce by half of actual coefficient correlation;

According to coefficient R _jiquantity N with every class customer complaint _iwhat calculate every class customer complaint estimates quantity T _ifor:

$T_i=\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{N}_i\&CenterDot;R_{\mathrm{ji}}/2;$

Preferably, can be further according to the average complaint probability C of every class customer complaint _ithat calculates every class customer complaint estimates quantity T _ifor:

$T_i=\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{R}_{\mathrm{ji}}/2\&CenterDot;N_i/C_i.$

If three,, owing to existing same type to complain particularly in the situation by same customer complaint in customer complaint in the past, client perception can sharply decline, and therefore needs according to complained the analysis that adds up in the past.

Be captured in the quantity S of the i class customer complaint being resolved in the t time period _i(t), obvious S _i(t) should be not more than the customer complaint quantity T collecting in this time period _i(t), i.e. 0 < S _i(t)≤T _i(t), the customer complaint of t+1 time period i class is estimated quantity and is:

$T_i(t+1)=N_i\left(t\right)\&CenterDot;(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)});$

Preferably, can be further according to the coefficient R between dissimilar customer complaint in the manner _jithe quantity of estimating of calculating every class customer complaint is:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{N}_i\&CenterDot;R_{\mathrm{ji}}/2;$

More preferably, in the manner, can also be further according to the average complaint probability C of every class customer complaint _ithe quantity of estimating of calculating every class customer complaint is:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{R}_{\mathrm{ji}}/2\&CenterDot;N_i/C_i.$

Also can be according to the quantity S of the i class customer complaint that obtains solving in this time period _i(t) the customer complaint quantity T, collecting in this time period _iand the average complaint probability C of every class customer complaint (t) _ithe quantity of estimating of calculating every class customer complaint is:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;N_i/C_i.$

In addition, be the accuracy and comprehensive that ensures to amplify, in above-mentioned three kinds of modes, average the rate of complaints C _i, coefficient R _jiall should regularly constantly update.

Customer complaint analytical method of the present invention, not only for some customer complaints of newly receiving, also be not limited to a small amount of customer complaint reflection problem, more considered some due to the network coverage less than and cannot complain or user's use business in some inconveniences or the not good content of perception situation about may not complain, by introducing amplification coefficient (as average the rate of complaints, quantity of coefficient correlation and the wherein class customer complaint that is resolved etc.) the customer complaint quantity having collected is amplified, realize the comprehensive situation that presents network quality by the complaint of certain customers by forming the situation of dissimilar customer complaint, the problem existing in can definite network, and then can be optimized network according to the demand of user awareness, improve the specific aim of the network optimization.

As shown in Figure 1, estimating after quantity of the every class customer complaint of acquisition, according to the data that obtain, network is optimized, concrete network optimization flow process is as follows:

Step 102, sorts the quantity of estimating of variety classes customer complaint, can sort according to the size of estimating quantity;

Step 104, problem and the region of complaining according to all types of user, obtain the current performance index value of certain entity in this region by testing equipment or network management, monitoring system, contrast existing network entity performance index value and the predefined performance index threshold value of operator, determine the performance index that do not meet the demands, need optimization;

Step 106, operation personnel divides definite parameter that affects performance index value according to previous experiences or device parameter function.

Step 108, determines the size that this parameter need to be adjusted;

Step 110, corresponding adjustment parameter;

Step 112, compares network entity performance index value and corresponding predefined performance index threshold value after adjusting, judges whether it meets predetermined requirement, if so, and execution step 114; If not, execution step 118;

Step 114, obtains current other performance index values of certain entity by testing equipment or network management, monitoring system;

Step 116, compares other network entity performance index values after adjusting and corresponding threshold value, judges whether it meets predetermined requirement, if so, shows that new capability desired value meets pre-provisioning request, uses later and adjusts rear parameter value, and optimizing process completes; If not, return to execution step 104;

Step 116, judges whether the parameter value after adjusting has exceeded predetermined range of parameter values, if so, returns to execution step 106; If not, return to execution step 108.

For example: complaining the customer complaint quantity that certain regional signal cannot access is 5, complaining probability is 25%, and according to above-mentioned analytical method one, the quantity of estimating of customer complaint is 5/0.25=20;

Whether access level lower than certain threshold according to this complaint value observation call completing rate and actual test afterwards, adjust access level threshold and neighbor cell configuration, ensure that user is accessible;

Finally whether test access level and call completing rate reach certain threshold again, and carry out real road test and see that access situation judges whether to meet the demands.

As shown in Figure 2, customer complaint analytical equipment embodiment of the present invention comprises:

Acquisition module 22, for gathering the quantity of every class customer complaint;

Statistical module 24, for adding up the average complaint probability of every class customer complaint;

Computing module 26, for the quantity of estimating that goes out every class customer complaint according to the average complaint probability calculation of the quantity of described every class customer complaint and every class customer complaint is: T _i=N _i/ C _i.

As shown in Figure 3, another embodiment of customer complaint analytical equipment of the present invention comprises:

Acquisition module 32, for gathering the quantity of every class customer complaint;

The first computing module 34, for calculating the coefficient correlation between dissimilar customer complaint;

The second computing module 36, estimates quantity T for what calculate every class customer complaint according to the quantity of described coefficient correlation and every class customer complaint _ifor:

Preferably, this device also comprises statistical module 38, for adding up the average complaint probability of every class customer complaint;

The second computing module 36, for further estimating quantity T according to the every class customer complaint of average complaint probability calculation of every class customer complaint _ifor:

As shown in Figure 4, customer complaint analytical equipment of the present invention again an embodiment comprise:

Acquisition module 42, for the wherein quantity of a class customer complaint that is captured in the quantity of every class customer complaint in certain hour section and is resolved within this time period;

The second computing module 44, what calculate the every class customer complaint of next time period for the quantity of a wherein class customer complaint that is resolved described in basis and the quantity of every class customer complaint estimates quantity T _ifor:

$T_i(t+1)=N_i\left(t\right)\&CenterDot;(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)}).$

Preferably, this device also comprises:

The first computing module 46, for calculating the coefficient correlation between dissimilar customer complaint;

The second computing module 44, for the further quantity of estimating according to the every class customer complaint of the Calculation of correlation factor between dissimilar customer complaint is:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{N}_i\&CenterDot;R_{\mathrm{ji}}/2.$

More preferably, this device also comprises:

Statistical module 48, for adding up the average complaint probability of every class customer complaint;

The second computing module 44, for according to the quantity of estimating of the every class customer complaint of average complaint probability calculation of every class customer complaint being further:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{R}_{\mathrm{ji}}/2\&CenterDot;N_i/C_i.$

Preferably, this device comprises: acquisition module 42, the second computing module 44 and statistical module 48,

The second computing module 44, for according to the quantity of estimating of the every class customer complaint of average complaint probability calculation of every class customer complaint being further:

wherein, C _ifor the average complaint probability of every class customer complaint.

The customer complaint analytical equipment of above-mentioned three embodiment, not only for some customer complaints of newly receiving, also be not limited to a small amount of customer complaint reflection problem, more considered some due to the network coverage less than and cannot complain or user's use business in some inconveniences or the not good content of perception situation about may not complain, by introducing amplification coefficient (as average the rate of complaints, quantity of coefficient correlation and the wherein class customer complaint that is resolved etc.) the customer complaint quantity having collected is amplified, realize the comprehensive situation that presents network quality by the complaint of certain customers by forming the situation of dissimilar customer complaint, the problem existing in can definite network, and then can be optimized network according to the demand of user awareness, improve the specific aim of the network optimization.

It should be noted that: above embodiment is only unrestricted in order to the present invention to be described, the present invention is also not limited in above-mentioned giving an example, and all do not depart from technical scheme and the improvement thereof of the spirit and scope of the present invention, and it all should be encompassed in claim scope of the present invention.

Claims (16)

1. a network optimized approach, is characterized in that, comprising:

Analyze the quantity of estimating of every class customer complaint;

The quantity of estimating to described every class customer complaint sorts;

According to the problem of every class customer complaint and region, obtain the current performance index value of certain entity in this region, contrast existing network entity performance index value and the predefined performance index threshold value of operator, determine the performance index that do not meet the demands, need optimization;

Determine the parameter of the described performance index of impact and the size that described parameter need to be adjusted;

The size that need to adjust according to described parameter is adjusted described parameter.

2. method according to claim 1, is characterized in that, the every class customer complaint of described analysis estimate quantity, comprising:

Gather the quantity of every class customer complaint;

Add up the average complaint probability of every class customer complaint;

The quantity of estimating that goes out every class customer complaint according to the average complaint probability calculation of the quantity of described every class customer complaint and every class customer complaint is:

T _i=N _i/ C _i, wherein, T _ibe the quantity of estimating of i class customer complaint, N _ifor the quantity of every class customer complaint of collecting, C _ifor the average complaint probability of every class customer complaint.

3. method according to claim 1, is characterized in that, the every class customer complaint of described analysis estimate quantity, comprising:

Gather the quantity of every class customer complaint;

Calculate the coefficient correlation between dissimilar customer complaint;

What calculate every class customer complaint according to the quantity of described coefficient correlation and every class customer complaint estimates quantity T _ifor:

wherein, R _jibe the coefficient correlation between the customer complaint of j class and the customer complaint of i class, N _ifor the quantity of every class customer complaint of collecting, M is the quantity of customer complaint classification.

4. method according to claim 3, is characterized in that, the every class customer complaint of described analysis estimate quantity, also comprise:

Further estimate quantity T according to the every class customer complaint of average complaint probability calculation of every class customer complaint _ifor:

wherein, C _ifor the average complaint probability of every class customer complaint.

5. method according to claim 1, is characterized in that, the every class customer complaint of described analysis estimate quantity, comprising:

Be captured in the quantity of every class customer complaint in certain hour section and the wherein quantity of a class customer complaint being resolved within this time period;

What calculate the every class customer complaint of next time period according to the described quantity of a wherein class customer complaint being resolved and the quantity of every class customer complaint estimates quantity T _ifor:

wherein, T _i(t) the customer complaint quantity for collecting in the t time period, S _i(t) quantity being resolved for i class customer complaint in the t time period, N _i(t) be the quantity of every class customer complaint of collecting in the t time period.

6. method according to claim 5, is characterized in that, the every class customer complaint of described analysis estimate quantity, also comprise:

According to the quantity of estimating of the every class customer complaint of the Calculation of correlation factor between dissimilar customer complaint be further:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{N}_i\&CenterDot;R_{\mathrm{ji}}/2,$ Wherein, R _jibe the coefficient correlation between the customer complaint of j class and the customer complaint of i class, N _ifor the quantity of every class customer complaint of collecting, M is the quantity of customer complaint classification.

7. method according to claim 6, is characterized in that, the every class customer complaint of described analysis estimate quantity, also comprise:

According to the quantity of estimating of the every class customer complaint of average complaint probability calculation of every class customer complaint be further:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{R}_{\mathrm{ji}}/2\&CenterDot;N_i/C_i,$ Wherein, C _ifor the average complaint probability of every class customer complaint.

8. method according to claim 5, is characterized in that, the every class customer complaint of described analysis estimate quantity, also comprise:

According to the quantity of estimating of the every class customer complaint of average complaint probability calculation of every class customer complaint be further:

wherein, C _ifor the average complaint probability of every class customer complaint, N _ifor the quantity of every class customer complaint of collecting.

9. a network optimization device, is characterized in that, comprises

For analyzing the device of estimating quantity of every class customer complaint;

For the device that quantity sorts of estimating to described every class customer complaint;

Be used for according to the problem of every class customer complaint and region, obtain the current performance index value of certain entity in this region, contrast existing network entity performance index value and the predefined performance index threshold value of operator, determine the device that does not meet the demands, needs the performance index of optimizing;

For determining the parameter of the described performance index of impact and the big or small device that described parameter need to be adjusted;

The device of described parameter being adjusted for the size that need to adjust according to described parameter.

10. device according to claim 9, is characterized in that, described is customer complaint analytical equipment for the device of estimating quantity of analyzing every class customer complaint, comprising:

Acquisition module, for gathering the quantity of every class customer complaint;

Statistical module, for adding up the average complaint probability of every class customer complaint;

Computing module, for the quantity of estimating that goes out every class customer complaint according to the average complaint probability calculation of the quantity of described every class customer complaint and every class customer complaint is:

T _i=N _i/ C _i, wherein, T _ibe the quantity of estimating of i class customer complaint, N _ifor the quantity of every class customer complaint of collecting, C _ifor the average complaint probability of every class customer complaint.

11. devices according to claim 9, is characterized in that, described is customer complaint analytical equipment for the device of estimating quantity of analyzing every class customer complaint, comprising:

Acquisition module, for gathering the quantity of every class customer complaint;

The first computing module, for calculating the coefficient correlation between dissimilar customer complaint;

The second computing module, estimates quantity T for what calculate every class customer complaint according to the quantity of described coefficient correlation and every class customer complaint _ifor:

wherein, R _jibe the coefficient correlation between the customer complaint of j class and the customer complaint of i class, N _ifor the quantity of every class customer complaint of collecting, M is the quantity of customer complaint classification.

12. devices according to claim 11, is characterized in that, described customer complaint analytical equipment, also comprises:

Statistical module, for adding up the average complaint probability of every class customer complaint;

Described the second computing module, for further estimating quantity T according to the every class customer complaint of average complaint probability calculation of every class customer complaint _ifor:

wherein, C _ifor the average complaint probability of every class customer complaint.

13. devices according to claim 9, is characterized in that, described is customer complaint analytical equipment for the device of estimating quantity of analyzing every class customer complaint, comprising:

The first acquisition module, for the wherein quantity of a class customer complaint that is captured in the quantity of every class customer complaint in certain hour section and is resolved within this time period;

The 3rd computing module, what calculate the every class customer complaint of next time period for the quantity of a wherein class customer complaint that is resolved described in basis and the quantity of every class customer complaint estimates quantity T _ifor:

wherein, S _i(t) quantity being resolved for i class customer complaint in the t time period, T _i(t) the customer complaint quantity for collecting in the t time period, N _i(t) be the quantity of every class customer complaint of collecting in the t time period.

14. devices according to claim 13, is characterized in that, described customer complaint analytical equipment also comprises:

The first computing module, for calculating the coefficient correlation between dissimilar customer complaint;

Described the 3rd computing module, for the further quantity of estimating according to the every class customer complaint of the Calculation of correlation factor between dissimilar customer complaint is:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{N}_i\&CenterDot;R_{\mathrm{ji}}/2,$ Wherein, R _jibe the coefficient correlation between the customer complaint of j class and the customer complaint of i class, N _ifor the quantity of every class customer complaint of collecting, M is the quantity of customer complaint classification.

15. devices according to claim 14, is characterized in that, described customer complaint analytical equipment also comprises:

Statistical module, for adding up the average complaint probability of every class customer complaint;

Described the 3rd computing module, for according to the quantity of estimating of the every class customer complaint of average complaint probability calculation of every class customer complaint being further:

$T_i(t+1)=(1+\underset{t}{\mathrm{\&Sigma;}}\frac{T_i\left(t\right)-S_i\left(t\right)}{T_i\left(t\right)})\&CenterDot;\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}{R}_{\mathrm{ji}}/2\&CenterDot;N_i/C_i,$ Wherein, C _ifor the average complaint probability of every class customer complaint.

16. devices according to claim 13, is characterized in that, described customer complaint analytical equipment also comprises:

Statistical module, for adding up the average complaint probability of every class customer complaint;

Described the 3rd computing module, for according to the quantity of estimating of the every class customer complaint of average complaint probability calculation of every class customer complaint being further:

wherein, C _ifor the average complaint probability of every class customer complaint, N _ifor the quantity of every class customer complaint of collecting.